[evlatests] Oct 15 pointing run

[Ken Sowinski]

Both EVLA antennas had their pointing model coefficients updated
on the afternoon of Oct 13 from the results of the Oct 12 pointing 
run.  The final post reconfiguration pointing run was on Oct 15
confirmed the changes and found several mistakes.

Antenna 14 was quite good except that the elevation collimation 
error had doubled because the correction was made with the wrong
sign.  Antenna 16 was not there at all.  Inspection of the pointing
parameters showed that I had entered the Az offset for 14 as that for
antenna 16 by mistake; this is an error of a degree and a half and
explains why 16 was not working on Friday.  All the mistakes have
now been corrected and visually checked.

For the record, the solution for antenna 14 was quite good.  the
chi square is 0.94 and the post-fit RMS 4.3 and 5.0 arc-sec in azimuth
and elevation repectively when refraction is fitted.  I see no evidence
for any problems with either the antenna or the software.  Some time
next week should be devoted to checking the inter-band collimations,

The lesson to be learned from this is that we should use a more
automatic method to update the pointing parameters.  Even two antennas
are too many to reliably modify by hand.  The system to do this has
been tested; we should make plans to begin using it and the new version
of peek by the next pointing run.